To compare the effects of two types of partial sleep deprivation (PSD) at the beginning (PSDBN) and the end (PSDEN) of the night on mood, cognitive performances, biomarkers of muscle damage, haematological status and antioxidant responses before and after repeated-sprint exercise in the post-lunch dip. Fourteen male athletes performed the Running-based Anaerobic Sprint Test following: (i) baseline normal sleep night, (ii) PSDBN, or (iii) PSDEN in a randomized and counter-balanced order. During each condition, participants performed simple and choice reaction time tests, the Profile of Mood States, subjective sleepiness, and the Running-based Anaerobic Sprint Test. Plasma biomarkers of muscle damage, total blood count, and antioxidant activities were measured at rest and after the repeated sprint in the three conditions. PSDEN decreased Pmax (p=0.008; d=1.12), Pmean (p=0.002; d=1.33) and Pmin (p=0.006; d=1.15), whilst PSDBN decreased Pmean (p=0.04; d=0.68) and Pmin (p=0.028; d=0.58), in comparison with baseline. PSDEN exerted stronger effects on Pmax (p=0.013; d=0.74) and Pmean (p=0.048; d=0.54) than PSDBN. Moreover, PSDEN increased subjective sleepiness (p<0.001; d=1.93), while PSDBN impaired choice reaction time (p<0.001, d=1.89). Both PSD types decreased resting glutathione peroxidase (p<0.001; d=5.43, d=3.86), and increased aspartate amino-transferase levels (p<0.001; d=1.36, d=1.37) respectively for PSDEN and PSDBN. PSDEN decreased repeated-sprint performances more than PSDBN in the post-lunch dip. This could be explained by the lowered mood and resting antioxidant status and the increased inflammatory profile after PSDEN. Repeated-sprint exercise resulted in greater inflammation after PSDEN, despite the decreased physical performance. The drop of resting antioxidant defence and haemoglobin concentration after PSDEN could explain the increased sleep drive at the post-lunch dip.